Papers › KartalOl: Transfer learning using deep neural network for iris segmentation and...

KartalOl: Transfer learning using deep neural network for iris segmentation and localization: New dataset for iris segmentation

9 Dec 2021arXiv:2112.05236archive 2025-07-28

Jalil Nourmohammadi Khiarak, Samaneh Salehi Nasab, Farhang Jaryani, Seyed Naeim Moafinejad, Rana Pourmohamad, Yasin Amini, Morteza Noshad

Iris segmentation and localization in unconstrained environments is challenging due to long distances, illumination variations, limited user cooperation, and moving subjects. To address this problem, we present a U-Net with a pre-trained MobileNetV2 deep neural network method. We employ the pre-trained weights given with MobileNetV2 for the ImageNet dataset and fine-tune it on the iris recognition and localization domain. Further, we have introduced a new dataset, called KartalOl, to better evaluate detectors in iris recognition scenarios. To provide domain adaptation, we fine-tune the MobileNetV2 model on the provided data for NIR-ISL 2021 from the CASIA-Iris-Asia, CASIA-Iris-M1, and CASIA-Iris-Africa and our dataset. We also augment the data by performing left-right flips, rotation, zoom, and brightness. We chose the binarization threshold for the binary masks by iterating over the images in the provided dataset. The proposed method is tested and trained in CASIA-Iris-Asia, CASIA-Iris-M1, CASIA-Iris-Africa, along the KartalOl dataset. The experimental results highlight that our method surpasses state-of-the-art methods on mobile-based benchmarks. The codes and evaluation results are publicly available at https://github.com/Jalilnkh/KartalOl-NIR-ISL2021031301.

Paper PDF Code

Code

jalilnkh/kartalol-nir-isl2021031301 officialmentioned in papermentioned on GitHubtf report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Binarization Domain Adaptation Iris Recognition Iris Segmentation Transfer Learning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

1x1 Convolution Average Pooling Batch Normalization Concatenated Skip Connection Convolution Depthwise Convolution Depthwise Separable Convolution Inverted Residual Block Max Pooling Pointwise Convolution ReLU U-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections